Cost Optimization for HCGaming
• There was two game SKU running in same structure, only the layout and configurations
was different.
• From previous disaster, customer turn on an hot-standing-by stack.
• Customer migrate DB from EC2 to Aurora, complained about the cost increased.
• Customer enabled the ASG for two main clusters.
• For OPS' convenience, customer migrated RabbitMQ from EC2 to AmazonMQ.
• Log files are piping into S3 and CloudWatch at the same time, since customer is learning
to build CloudWatch Alarm, hook with SNS and Slack for OPS reason.
• There was no complex logic in Analytics Services, those machines was running 7x24 but
mostly idled.
• Customer does not really understand why there is a queue before Analytics service and
then after, since they now running daily and weekly report.
• Data Ingest Services was running licensed package from other vendor, simply docker but
running stand-alone for each server.